Most motor control studies consider how the CNS controls task level variables, examining, for example, how the CNS produces the joint torques necessary to achieve behaviors such as locomotion. In this context, it is the set of torques produced by a muscle that determines its activation by the CNS. However, this focus on task performance ignores the control of another critical set of variables, those characterizing the state of internal joint structures such as ligaments and articular cartilage (i.e. ligament strainsor bone contact forces). Failure to regulate these internal joint variables can have significant consequences to health both in the short term (e.g. ligament rupture, joint dislocation) and in the long term (e.g. chronic joint pain, arthritis). The CNS should therefore consider both task performance and internal joint variables when determining muscle activations. How internal joint variables might be incorporated into motor control strategies, however, is poorly understood. The overall goal of the experiments described in this proposal is to evaluate these issues, examining the control of internal joint variables by the CNS. We will examine these issues using an animal model, focusing on the control of the knee joint by quadriceps muscles in the rat. The specific anatomy of the rat knee allows for a clear separation between the effects of quadriceps muscles on task performance variables (joint torques) and internal joint variables (mediolateral patellar forces). Using this model we can therefore make strong predictions about how the control of internal joint variables should be reflected in muscle activations across a range of behavioral conditions. We will perform three sets of related experiments. In Aim 1 we will characterize the mechanical actions of quadriceps muscles on task performance and internal joint variables. We hypothesize that quadriceps muscles will produce similar knee joint torques but distinct mediolateral patellar forces. In Aim 2, we will examine whether the neural control of quadriceps reflects the regulation of internal joint variables. We first hypothesize that in intact animals, the correlation in the variability of EMGs reflects the balancing of mediolateral patellar forces. Further, we hypothesize that following selective muscle paralysis or perturbations of patellar forces, long term adaptations in muscle activations will improve the control of internal joint variables. In Aim 3 we will examine the role of joint afferents in the control of internal jont variables. We hypothesize that joint afferents are not used for rapid feedback control of muscle activations but are used to guide long term adaptations of muscle activations following perturbations to internal joint variables. These experiments provide a systematic analysis of the role of internal joint variables in the neural control of behavior, using a range of techniques in conceptually simple and tractable experimental model. The results of these experiments have the potential to significantly impact motor control, both in our basic understanding of motor control and in clinical applications that seek to restore function after injury.